Use built-in method to fix HTML link resolution #2273

wodin · 2025-05-02T08:41:35Z

I went through the Git history for the HTML app source to find all of the commits where the ensureAbsoluteUrl function was modified, and the related issues/discussions. As far as I can see this PR will not cause any regressions.

I also created a test script to verify that the built-in resolve() method works as expected. The script also shows that there are various cases that the current custom implementation does not handle correctly.

String ensureAbsoluteUrl(String ambiguousUrl, Uri referenceAbsoluteUrl) {
  if (ambiguousUrl.startsWith('//')) {
    ambiguousUrl = '${referenceAbsoluteUrl.scheme}:$ambiguousUrl';
  }
  try {
    Uri.parse(ambiguousUrl).origin;
    return ambiguousUrl;
  } catch (err) {
    // is relative
  }
  var currPathSegments = referenceAbsoluteUrl.path
      .split('/')
      .where((element) => element.trim().isNotEmpty)
      .toList();
  String absoluteUrl;
  if (ambiguousUrl.startsWith('/')) {
    absoluteUrl = '${referenceAbsoluteUrl.origin}$ambiguousUrl';
  } else if (currPathSegments.isEmpty) {
    absoluteUrl = '${referenceAbsoluteUrl.origin}/$ambiguousUrl';
  } else if (ambiguousUrl.split('/').where((e) => e.isNotEmpty).length == 1) {
    absoluteUrl =
        '${referenceAbsoluteUrl.origin}/${currPathSegments.join('/')}/$ambiguousUrl';
  } else {
    absoluteUrl =
        '${referenceAbsoluteUrl.origin}/${currPathSegments.sublist(0, currPathSegments.length - (currPathSegments.last.contains('.') ? 1 : 0)).join('/')}/$ambiguousUrl';
  }
  return Uri.parse(absoluteUrl).toString();
}

void compareApproaches(String ref, String base) {
  var baseUri = Uri.parse(base);
  var custom = ensureAbsoluteUrl(ref, baseUri);
  var builtIn = baseUri.resolve(ref).toString();

  if (custom == builtIn) {
    print('PASS - ${base}, ${ref}');
  } else {
    print('FAIL - ${base}, ${ref} => ${custom}');
  }
}

void main() {
  compareApproaches('test.apk', 'https://www.example.com/download.html');                   // FAIL
  compareApproaches('//www.example.com/test.apk', 'https://www.example.com/download.html'); // PASS
  compareApproaches('//test.apk', 'https://www.example.com/download.html');                 // PASS
  compareApproaches('path//test.apk', 'https://www.example.com/download.html');             // FAIL
  compareApproaches('./test.apk', 'https://www.example.com/download.html');                 // FAIL
  compareApproaches('../test.apk', 'https://www.example.com/x/download.html');              // PASS
  compareApproaches('../test.apk', 'https://www.example.com/download.html');                // PASS
  compareApproaches('../test.apk', 'https://www.example.com/x/README');                     // FAIL
  compareApproaches('../test.apk', 'https://www.example.com/README');                       // PASS
  compareApproaches('../test.apk', 'https://www.example.com/x/dl.dir/');                    // FAIL
  compareApproaches('../test.apk', 'https://www.example.com/dl.dir/');                      // PASS

  print('#848');
  compareApproaches('14.5a6/', 'https://dist.torproject.org/torbrowser/');                  // PASS
  compareApproaches('tor-browser-android-aarch64-14.5a6.apk', 'https://dist.torproject.org/torbrowser/14.5a6/');        // PASS

  print('#989');
  compareApproaches('download/Free42Android.apk', 'https://thomasokken.com/free42/');       // PASS

  print('#1253');
  compareApproaches('2.2.0/', 'https://secure.nic.cz/files/datove_schranky/mobile-datovka/');                   // PASS
  compareApproaches('mobile-datovka-2.2.0-android-arm64.apk', 'https://datovka.nic.cz/mobile-datovka/2.2.0/');  // PASS

  print('#1259');
  compareApproaches('/stable/1.21.0/', 'https://buildbot.libretro.com/stable');                                                 // PASS
  compareApproaches('/stable/1.21.0/', 'https://buildbot.libretro.com/stable/');                                                // PASS
  compareApproaches('/stable/1.21.0/android/', 'https://buildbot.libretro.com/stable/1.21.0/');                                 // PASS
  compareApproaches('/stable/1.21.0/android/RetroArch_aarch64.apk', 'https://buildbot.libretro.com/stable/1.21.0/android/');    // PASS

  print('#1937');
  compareApproaches('com.brave.browser_1.71.118.apk', 'https://apk.example.com');           // PASS
  compareApproaches('com.brave.browser_1.71.118.apk', 'https://apk.example.com/');          // PASS
  compareApproaches('com.brave.browser_1.71.118.apk', 'https://apk.example.com/path/');     // PASS

  print('#1998');
  // Looks like the fix was for links starting with '//host', but Signal/TopoDroid don't currently do that
  compareApproaches('//updates.signal.org/test.apk', 'https://updates.signal.org/android/latest.json');             // PASS
  compareApproaches('https://updates.signal.org/android/Signal-Android-website-prod-universal-release-7.40.2.apk',
                    'https://updates.signal.org/android/latest.json');                                              // PASS
  compareApproaches('topodroid_apk/TopoDroidX-6.3.20-35.apk',
                    'http://marcocorvi.altervista.org/caving/speleoapps/speleoapks/TopoDroidApks.html');            // PASS

  var base = 'http://a/b/c/d;p?q';
  var baseUri = Uri.parse(base);

  // Test cases from RFC 3986
  // https://datatracker.ietf.org/doc/html/rfc3986#section-5.4
  var testCases = [
    // reference URL, expected result
    ['g:h', 'g:h'],                             //  0
    ['g', 'http://a/b/c/g'],                    //  1
    ['./g', 'http://a/b/c/g'],                  //  2
    ['g/', 'http://a/b/c/g/'],                  //  3
    ['/g', 'http://a/g'],                       //  4
    ['//g', 'http://g'],                        //  5
    ['?y', 'http://a/b/c/d;p?y'],               //  6
    ['g?y', 'http://a/b/c/g?y'],                //  7
    ['#s', 'http://a/b/c/d;p?q#s'],             //  8
    ['g#s', 'http://a/b/c/g#s'],                //  9
    ['g?y#s', 'http://a/b/c/g?y#s'],            // 10
    [';x', 'http://a/b/c/;x'],                  // 11
    ['g;x', 'http://a/b/c/g;x'],                // 12
    ['g;x?y#s', 'http://a/b/c/g;x?y#s'],        // 13
    ['', 'http://a/b/c/d;p?q'],                 // 14
    ['.', 'http://a/b/c/'],                     // 15
    ['./', 'http://a/b/c/'],                    // 16
    ['..', 'http://a/b/'],                      // 17
    ['../', 'http://a/b/'],                     // 18
    ['../g', 'http://a/b/g'],                   // 19
    ['../..', 'http://a/'],                     // 20
    ['../../', 'http://a/'],                    // 21
    ['../../g', 'http://a/g'],                  // 22

    ['../../../g', 'http://a/g'],               // 23
    ['../../../../g', 'http://a/g'],            // 24

    ['/./g', 'http://a/g'],                     // 25
    ['/../g', 'http://a/g'],                    // 26
    ['g.', 'http://a/b/c/g.'],                  // 27
    ['.g', 'http://a/b/c/.g'],                  // 28
    ['g..', 'http://a/b/c/g..'],                // 29
    ['..g', 'http://a/b/c/..g'],                // 30

    ['./../g', 'http://a/b/g'],                 // 31
    ['./g/.', 'http://a/b/c/g/'],               // 32
    ['g/./h', 'http://a/b/c/g/h'],              // 33
    ['g/../h', 'http://a/b/c/h'],               // 34
    ['g;x=1/./y', 'http://a/b/c/g;x=1/y'],      // 35
    ['g;x=1/../y', 'http://a/b/c/y'],           // 36

    ['g?y/./x', 'http://a/b/c/g?y/./x'],        // 37
    ['g?y/../x', 'http://a/b/c/g?y/../x'],      // 38
    ['g#s/./x', 'http://a/b/c/g#s/./x'],        // 39
    ['g#s/../x', 'http://a/b/c/g#s/../x'],      // 40

    ['http:g', 'http:g'],                       // 41
  ];

  print('');
  print('Test cases from RFC 3986');
  for (var i = 0; i < testCases.length; i++) {
    var testCase = testCases[i];
    var ref = testCase[0];
    var expected = testCase[1];

    // var result = baseUri.resolve(ref).toString();
    var result = ensureAbsoluteUrl(ref, baseUri);
    if (result == expected) {
      print('PASS - ${i}');
    } else {
      print('FAIL - ${i} - "${base}", "${ref}", "${expected}" != "${result}"');
    }
  }
}

Results

FAIL - https://www.example.com/download.html, test.apk => https://www.example.com/download.html/test.apk
PASS - https://www.example.com/download.html, //www.example.com/test.apk
PASS - https://www.example.com/download.html, //test.apk
FAIL - https://www.example.com/download.html, path//test.apk => https://www.example.com//path//test.apk
FAIL - https://www.example.com/download.html, ./test.apk => https://www.example.com//test.apk
PASS - https://www.example.com/x/download.html, ../test.apk
PASS - https://www.example.com/download.html, ../test.apk
FAIL - https://www.example.com/x/README, ../test.apk => https://www.example.com/x/test.apk
PASS - https://www.example.com/README, ../test.apk
FAIL - https://www.example.com/x/dl.dir/, ../test.apk => https://www.example.com/test.apk
PASS - https://www.example.com/dl.dir/, ../test.apk

#848

PASS - https://dist.torproject.org/torbrowser/, 14.5a6/
PASS - https://dist.torproject.org/torbrowser/14.5a6/, tor-browser-android-aarch64-14.5a6.apk

#989

PASS - https://thomasokken.com/free42/, download/Free42Android.apk

#1253

PASS - https://secure.nic.cz/files/datove_schranky/mobile-datovka/, 2.2.0/
PASS - https://datovka.nic.cz/mobile-datovka/2.2.0/, mobile-datovka-2.2.0-android-arm64.apk

#1259

PASS - https://buildbot.libretro.com/stable, /stable/1.21.0/
PASS - https://buildbot.libretro.com/stable/, /stable/1.21.0/
PASS - https://buildbot.libretro.com/stable/1.21.0/, /stable/1.21.0/android/
PASS - https://buildbot.libretro.com/stable/1.21.0/android/, /stable/1.21.0/android/RetroArch_aarch64.apk

#1937

PASS - https://apk.example.com, com.brave.browser_1.71.118.apk
PASS - https://apk.example.com/, com.brave.browser_1.71.118.apk
PASS - https://apk.example.com/path/, com.brave.browser_1.71.118.apk

#1998

PASS - https://updates.signal.org/android/latest.json, //updates.signal.org/test.apk
PASS - https://updates.signal.org/android/latest.json, https://updates.signal.org/android/Signal-Android-website-prod-universal-release-7.40.2.apk
PASS - http://marcocorvi.altervista.org/caving/speleoapps/speleoapks/TopoDroidApks.html, topodroid_apk/TopoDroidX-6.3.20-35.apk

Test cases from RFC 3986

FAIL - 0 - "http://a/b/c/d;p?q", "g:h", "g:h" != "http://a/b/c/d;p/g:h"
FAIL - 1 - "http://a/b/c/d;p?q", "g", "http://a/b/c/g" != "http://a/b/c/d;p/g"
FAIL - 2 - "http://a/b/c/d;p?q", "./g", "http://a/b/c/g" != "http://a/b/c/d;p/g"
FAIL - 3 - "http://a/b/c/d;p?q", "g/", "http://a/b/c/g/" != "http://a/b/c/d;p/g/"
PASS - 4
PASS - 5
FAIL - 6 - "http://a/b/c/d;p?q", "?y", "http://a/b/c/d;p?y" != "http://a/b/c/d;p/?y"
FAIL - 7 - "http://a/b/c/d;p?q", "g?y", "http://a/b/c/g?y" != "http://a/b/c/d;p/g?y"
FAIL - 8 - "http://a/b/c/d;p?q", "#s", "http://a/b/c/d;p?q#s" != "http://a/b/c/d;p/#s"
FAIL - 9 - "http://a/b/c/d;p?q", "g#s", "http://a/b/c/g#s" != "http://a/b/c/d;p/g#s"
FAIL - 10 - "http://a/b/c/d;p?q", "g?y#s", "http://a/b/c/g?y#s" != "http://a/b/c/d;p/g?y#s"
FAIL - 11 - "http://a/b/c/d;p?q", ";x", "http://a/b/c/;x" != "http://a/b/c/d;p/;x"
FAIL - 12 - "http://a/b/c/d;p?q", "g;x", "http://a/b/c/g;x" != "http://a/b/c/d;p/g;x"
FAIL - 13 - "http://a/b/c/d;p?q", "g;x?y#s", "http://a/b/c/g;x?y#s" != "http://a/b/c/d;p/g;x?y#s"
FAIL - 14 - "http://a/b/c/d;p?q", "", "http://a/b/c/d;p?q" != "http://a/b/c/d;p/"
FAIL - 15 - "http://a/b/c/d;p?q", ".", "http://a/b/c/" != "http://a/b/c/d;p/"
FAIL - 16 - "http://a/b/c/d;p?q", "./", "http://a/b/c/" != "http://a/b/c/d;p/"
FAIL - 17 - "http://a/b/c/d;p?q", "..", "http://a/b/" != "http://a/b/c/"
FAIL - 18 - "http://a/b/c/d;p?q", "../", "http://a/b/" != "http://a/b/c/"
FAIL - 19 - "http://a/b/c/d;p?q", "../g", "http://a/b/g" != "http://a/b/c/g"
FAIL - 20 - "http://a/b/c/d;p?q", "../..", "http://a/" != "http://a/b/"
FAIL - 21 - "http://a/b/c/d;p?q", "../../", "http://a/" != "http://a/b/"
FAIL - 22 - "http://a/b/c/d;p?q", "../../g", "http://a/g" != "http://a/b/g"
PASS - 23
PASS - 24
PASS - 25
PASS - 26
FAIL - 27 - "http://a/b/c/d;p?q", "g.", "http://a/b/c/g." != "http://a/b/c/d;p/g."
FAIL - 28 - "http://a/b/c/d;p?q", ".g", "http://a/b/c/.g" != "http://a/b/c/d;p/.g"
FAIL - 29 - "http://a/b/c/d;p?q", "g..", "http://a/b/c/g.." != "http://a/b/c/d;p/g.."
FAIL - 30 - "http://a/b/c/d;p?q", "..g", "http://a/b/c/..g" != "http://a/b/c/d;p/..g"
FAIL - 31 - "http://a/b/c/d;p?q", "./../g", "http://a/b/g" != "http://a/b/c/g"
FAIL - 32 - "http://a/b/c/d;p?q", "./g/.", "http://a/b/c/g/" != "http://a/b/c/d;p/g/"
FAIL - 33 - "http://a/b/c/d;p?q", "g/./h", "http://a/b/c/g/h" != "http://a/b/c/d;p/g/h"
FAIL - 34 - "http://a/b/c/d;p?q", "g/../h", "http://a/b/c/h" != "http://a/b/c/d;p/h"
FAIL - 35 - "http://a/b/c/d;p?q", "g;x=1/./y", "http://a/b/c/g;x=1/y" != "http://a/b/c/d;p/g;x=1/y"
FAIL - 36 - "http://a/b/c/d;p?q", "g;x=1/../y", "http://a/b/c/y" != "http://a/b/c/d;p/y"
FAIL - 37 - "http://a/b/c/d;p?q", "g?y/./x", "http://a/b/c/g?y/./x" != "http://a/b/c/d;p/g?y/./x"
FAIL - 38 - "http://a/b/c/d;p?q", "g?y/../x", "http://a/b/c/g?y/../x" != "http://a/b/c/d;p/g?y/../x"
FAIL - 39 - "http://a/b/c/d;p?q", "g#s/./x", "http://a/b/c/g#s/./x" != "http://a/b/c/d;p/g#s/./x"
FAIL - 40 - "http://a/b/c/d;p?q", "g#s/../x", "http://a/b/c/g#s/../x" != "http://a/b/c/d;p/g#s/../x"
FAIL - 41 - "http://a/b/c/d;p?q", "http:g", "http:g" != "http://a/b/c/d;p/http:g"

wodin · 2025-05-05T09:03:44Z

I have finally managed to build a version of Obtainium with my changes applied and run it in an emulator. I can confirm that this PR fixes the problem for me.

ImranR98 · 2025-05-05T18:52:40Z

Thanks, this is very thorough! I had no idea there was a built in function.

Use built-in method to fix HTML link resolution

2f54de6

wodin mentioned this pull request May 5, 2025

When HTML source URL ends with /download.html, Obtainium cannot find the APK #2272

Closed

ImranR98 merged commit 20d3c41 into ImranR98:main May 5, 2025

wodin deleted the use-built-in-url-resolver branch May 5, 2025 21:30

kratz00 mentioned this pull request May 19, 2025

Fix com.collabora.libreoffice.* URL values. ImranR98/apps.obtainium.imranr.dev#682

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Use built-in method to fix HTML link resolution #2273

Use built-in method to fix HTML link resolution #2273

Uh oh!

wodin commented May 2, 2025 •

edited

Loading

Uh oh!

wodin commented May 5, 2025

Uh oh!

ImranR98 commented May 5, 2025

Uh oh!

Uh oh!

Use built-in method to fix HTML link resolution #2273

Use built-in method to fix HTML link resolution #2273

Uh oh!

Conversation

wodin commented May 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Results

Test cases from RFC 3986

Uh oh!

wodin commented May 5, 2025

Uh oh!

ImranR98 commented May 5, 2025

Uh oh!

Uh oh!

wodin commented May 2, 2025 •

edited

Loading